Symbolic Neural Ordinary Differential Equations
Xin Li, Chengli Zhao, Xue Zhang, Xiaojun Duan · 2025
Read Symbolic Neural Ordinary Differential Equations by Xin Li, Chengli Zhao, Xue Zhang, Xiaojun Duan (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-08059.txt.
Description
Xin Li, Chengli Zhao, Xue Zhang, and Xiaojun Duan present a learning system that reads the rules behind moving systems. Many natural and built systems change over time and are summed up by equations. Their 2025 system, called SNODEs, mixes symbol-based networks with deep learning. The aim is a model that fits the data, stays clear to read, and holds up on new cases.
Training runs in three stages. First the symbol network learns from the flow of the data. Then it is tuned with a depth-based learning method. Last, a plain network mops up what is left. Tests on sample systems show the mix beats older methods on reach and clarity. The authors see uses in control, forecasts, and the hunt for new equations.
Catalog record
| CREATOR | Xin Li, Chengli Zhao, Xue Zhang, Xiaojun Duan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-08059